Chaffing and Winnowing: Confidentiality without Encryption. ⇒
24 December 2005, mid-afternoon
This is a really neat idea.
This is a post from my link log: If you click the title of this post you will be taken the web page I am discussing.
24 December 2005, mid-afternoon
This is a really neat idea.
This is a post from my link log: If you click the title of this post you will be taken the web page I am discussing.
I might have missed it, but what guarantees no collisions with the chaffing process? If a MAC collision were to occur and all-in-one packing was used, then you’ve got a useless collection of packets right?
by Haran on December 24 2005, 5:33 pm #
you’re right, but they do suggest picking the chaffspace to be large enough to make collision astronomically improbable.
if they’re doing packaging then the rx code would automatically detect a fault and discard the whole package and (presumably) call for a retransmit…the same thing that TCP (or etc.) would do anyhow.
the downside being that you have to ask for a huge chunk (whole package) all over again, but this is balanced by its unlikelihood.
I thought the paper was interesting, but all the “see, it’s not encryption and it’s exportable because all you’re “really” doing is adding authentication, which is technically legal, tee hee” fluff was a little tiresome.
by Weiguo on December 25 2005, 8:04 pm #